Environ. Sci. Technol. 2006, 40, 3313-3319
Nitrogen Removal in a Small Constructed Wetland: An Isotope Mass Balance Approach MIRIAM REINHARDT,* BEAT MU ¨ LLER, R E N EÄ G A¨ C H T E R , A N D BERNHARD WEHRLI Swiss Federal Institute of Aquatic Science and Technology (Eawag) and Swiss Federal Institute of Technology (ETH), CH-6047 Kastanienbaum, Switzerland
The nitrogen (N) removal potential of constructed wetlands is increasingly used to lower the N load from agricultural nonpoint sources to inland and coastal waters. To determine the removal efficiency and key factors limiting wetland N removal, N fluxes were studied in a small constructed wetland in Central Switzerland. With an isotope mass balance approach integrating the natural isotope signature of nitrate (NO3-), ammonium (NH4+), and particulate nitrogen (PN), the N transformations such as assimilation, mineralization, nitrification, and denitrification were quantified. On average, the wetland removed 45 g m-2 yr-1 N during the studied 2.5 years, corresponding to a removal efficiency of 27%. Denitrification contributed 94% to the N removal, while only 6% of the removed N accumulated in the sediments. Denitrification was most efficient during periods with an oxic water column overlying anoxic sediments, as NH4+ released during mineralization of sediment organic matter was completely nitrified and subsequently denitrified at the sediment-water interface. During water column anoxia, NH4+ accumulated in the water and fueled assimilation by duckweed and internal recycling. The NO3-N isotope signature in the wetland mainly reflected the mineralization of sediment organic matter and subsequent nitrification, while denitrification at the sediment-water interface produced no fractionation.
Introduction Wastewater discharge and the application of inorganic N fertilizer and liquid manure in excess of plant demands augment the N transfer to aquatic ecosystems worldwide (1, 2). Concerns over coastal eutrophication (3) has triggered political response such as the recommendation of the Paris Commission to reduce N inputs to the tributaries of the North East Atlantic by 50% (4). Improvements in wastewater treatment were achieved rapidly, but lowering agricultural nonpoint source pollution has proven to be more difficult and tedious. Natural and constructed wetlands have the potential to decrease the agricultural N discharge (5). Thus, recently in many countries in Europe and the United States, wetlands, riparian buffer strips, and ponds have been restored or artificially constructed at the interface between cultivated land and receiving waters. Fed with drainage water, storm* Corresponding author. Telephone: +41 41 349 21 53. Fax: +41 41 349 21 68. E-mail:
[email protected]. 10.1021/es052393d CCC: $33.50 Published on Web 04/13/2006
2006 American Chemical Society
water, or river water, these mostly small, low-maintenance “treatment plants” remove part of their N load as they trap particulate matter, produce and accumulate biomass, and eliminate N by denitrification (6). To which degree individual N transformations contribute to N removal and which factors limit these transformations are key questions for improving the removal efficiency of wetlands. The isotopic signature of N species reflects their origin (7) and allows the study of N processing. Microbial N transformations discriminate against the heavy N isotope, leaving the substrate 15N enriched and the product 15N depleted. In field studies determining denitrification in riparian buffer zones, decreasing NO3- concentrations have been successfully linked to an increasing accumulation of 15N in NO - (8). In wetlands, however, the quantitative 3 assessment of denitrification based on simple measurements of the isotopic NO3- composition has mostly failed, because additional fractionating processes were neglected or could not be assessed (9). In the present work, we studied the fluxes of all significant N species entering and leaving a small constructed wetland in Central Switzerland and investigated the natural isotopic composition not only of NO3-, but also of NH4+, of the biomass, and of the sediment particulate nitrogen in the wetland. Our objectives were (i) to investigate whether the relevant N transformations, in particular denitrification, nitrification, and mineralization, can be quantified with an isotope mass balance approach; (ii) to assess the relative importance of the processes responsible for the long-term N removal, and (iii) to identify key factors controlling the N transformations in wetlands.
Materials and Methods Study Site. Wetland “Boden“ is located in the Central Swiss Plateau at 540 m above sea level. Precipitation averages 1250 mm yr-1, and the mean annual temperature is approximately 10 °C. The predominantly loamy, water-logged soil of the watershed is drained with subsurface drainage pipes and cultivated as grassland. From the end of February to midNovember, liquid manure is frequently spread on the grassland. In winter 2001 and 2002, the surface-flow wetland was constructed at the interface between two subsurface drainage systems and the small receiving river “Waldbach”, a tributary of Lake Sempach. The wetland’s watershed spans 8.4 ha of slightly sloped grassland. The wetland comprises a total water surface area of 720 m2 with a volume of 700 m3 and a maximum depth of 3 m. Three separate basins were excavated, two small and flat sedimentation basins (20 m3 and 80 m3) and one deep pond (600 m3) (for a map, see Supporting Information). Both sedimentation basins, each fed by a separate drainage pipe (inlets I and II), discharge into the main pond. Inlet II delivers only soil drainage water, whereas inlet I additionally discharges surface water from the nearby farm yard. The wetland’s outlet pipe is embedded in the surrounding dam 0.6 m below its top edge. In summer 2003 and 2004 the water surface was covered with duckweed (Lemna minor L.). Hydrological Monitoring. From May 2002 to October 2004, the water discharge at inlet I and the outlet was measured with gaged V-notch weirs (ATM/N water level gage, STS, Sirnach, Switzerland). The water discharge of inlet II was measured with an inductive flow meter (Promag 53W, Endress+Hauser, Reinach, Switzerland). In the main pond the water level was monitored with an ATM/N gage. Readings of flow rate and water level were recorded every 10 min with VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3313
a CR10 data logger (Campbell Scientific, Logan, UT). Every 20 min, water temperature was measured 15 cm below the water surface (Minilog12, VEMCO, Shad Bay, Canada). Sampling. In 2002, water samples of inlet I, inlet II, and the outlet were collected manually every second day. In 2003 and 2004, two automatic samplers (6712, ISCO, Lincoln, NE; Manning 4900, TN Technologies, Round Rock, TX) sampled inlet I and the outlet whenever distinct amounts of water (inlet I: 4.3 m3; outlet: 8.6 m3) had passed through. A composite sample was made from 20 and 10 spot samples at the inlet I and the outlet, respectively. As inlet II contributed only 13% to the total water load (see below), its sampling was discontinued in 2003, but restarted in 2004 for isotope analysis. For 2003, its N load was estimated from the measured water discharge in 2003 and median concentrations of the different N species observed in 2002 and 2004. To study the nitrogen isotopic signature, inlet I, inlet II and the outlet were sampled manually at least once per month from March to October 2004. In addition, monthly samples of the water column, the duckweed and the sediment trap (radius: 7.1 cm, height: 50 cm) installed at the deepest site of the main pond were collected. Concurrently, oxygen and temperature profiles were measured in situ at approximately 10 a.m. (CellOx325, WTW, Weilheim, Germany). From the heat balance of the wetland the turbulent diffusion coefficient (10) was calculated. Sediment cores were collected in March 2003 and November 2004. Sediment and rinsed duckweed were freeze-dried and ground. Chemical Analysis. Water samples were filtered through 0.45-µm cellulose acetate filters and stored at 5 °C until analysis. NO3- and NH4+ concentrations were determined with standard photometric techniques (11). Total nitrogen (TN) and dissolved nitrogen (DN) of the water samples and TN of duckweed and sediment samples were measured after alkaline persulfate digestion (12). Particulate nitrogen was calculated as the difference between TN and DN. Dissolved (mainly) organic nitrogen (DON) resulted from the subtraction of NO3- and NH4+ from DN. For mass balance considerations PN was subdivided into allochthonous PN (PNal) and into autochthonous PN (PNau). All concentrations are given as g N m-3. Nitrogen isotope ratios of samples are reported in standard δ notation (‰) relative to atmospheric N2. The degree of fractionation during a reaction is expressed as enrichment factor (‰), which can be approximated as the difference of the δ15N of the product and the δ15N of the substrate (13). For isotope analysis of 15NNO3 and 15NNH4, samples were prepared according to Silva et al. (14) and Lehmann et al. (15), respectively. NO3- was collected on anion-exchange resins (AG 1-X8, 100-200 mesh size; Bio-Rad, Hercules, CA), extracted afterward and precipitated as AgNO3. NH4+ was collected on cation-exchange resins (AG 50W-X8, 100-200 mesh size, Bio-Rad, Hercules, CA), which were subsequently dried, ground and homogenized. AgNO3, NH4+-containing resin, samples of duckweed, sediment trap material, and sediment were combusted in an elemental analyzer (CNS 2500, Carlo Erba, Milan, Italy) and continuously transferred into an isotope ratio mass spectrometer (IsoPrime Stable Isotope Ratio MS, GV instruments, Manchester, UK). The reproducibility of the δ15N measurement was (0.2‰. Mass Balance Calculations. Denitrification (deni) and nitrification (nitr) were calculated combining the NO3- mass balance (eq 1) and the mass balance of its isotopic composition (eq 2). The change of the NO3- stock in the wetland’s water column during the investigated period (dNO3-/dt) equals the difference between NO3- sources, i.e., NO3- input (in) and nitrification, and NO3- sinks, i.e., NO3- output (out), 3314
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 10, 2006
TABLE 1. Nitrogen Balance of Wetland Boden (05/2002-10/2004) TN NO3--N NH4+-N DON PN
input (kg)
output (kg)
removal (%)
301 143 15 100 43
220 101 16 78 25
27 29 -1 22 42
denitrification, and uptake of NO3- into PNau (up).
dNO3 ) in + nitr - out - deni - up dt
(1)
NO3- fluxes at the inlet and outlet were calculated by multiplying the average time or flow proportional NO3concentrations by the water discharge integrated over the corresponding period. Uptake of NO3- into PNau includes production of plankton PNplankton, duckweed PNduckweed, and autochthonous sediment trap material PNtrap,au. Allochthonous PN was assumed to settle completely within the wetland. N2O production during nitrification and denitrification was disregarded in the mass balance as it usually amounts to less than 0.5% of the N2 flux (16). Analogous to eq 1, the mass balance for the isotopic composition of NO3- was calculated as
dNO3‚δ15NNO3 ) in‚δ15NNO3,in + nitr‚δ15NNO3,nitr dt out‚δ15NNO3,out - deni‚δ15NN2 - up‚δ15NPNau (2) The δ15N measured for duckweed was also applied to plankton, whose isotopic composition was not analytically determined. For δ15NN2 and δ15NNO3,nitr see discussion. Laboratory Experiments. Potential denitrification was measured in duplicate laboratory experiments recording the NO3- decrease from the water overlying two undisturbed sediment cores (radius: 2.9 cm, water column height: 21 cm). In three successive series, NO3- was added to the stirred water column to a final concentration of approximately 1.90 mg L-1, 4.98 mg L-1 and 15.60 mg L-1. Sediment cores were incubated at 9 °C in the dark and sampled after 6, 20, 30, 44 h and subsequently every 24 h following each spike.
Results Hydrologic Regime. Average hydraulic loading based on the wetland’s surface area was 150 mm d-1 in 2002, but only 50 mm d-1 in 2003 and 2004 due to exceptionally dry weather. Inlet I contributed 87% and inlet II 13% to the total water discharge. The outlet pipe discharged 81% of the water from the wetland, while 19% was lost as downward seepage. Nitrogen Balance. Average TN input to the wetland was 167 g m-2 yr-1 with respect to the wetland surface. Nitrate was the main N species in the drainage inlets of the wetland and contributed 48% of the TN load (Table 1). Median NO3-N concentration of inlet I and inlet II were 1.91 mg L-1 and 3.04 mg L-1, respectively. Up to 12.9 mg L-1 were detected in both inlets during the first intense rainfalls after the long dry period in summer 2003. During 2004, the NO3- loading, calculated from the water discharge (Figure 1a) and the NO3- concentration (Figure 1c), was high from January to June (292 mg m-2 d-1), but decreased subsequently due to the dry weather from July to October (13 mg m-2 d-1) (Table 2). δ15NNO3 of inlets I and II ranged from 13.1 to 24.4‰ and were typical for manure-derived δ15NNO3 (17); δ15NNO3 of the outlet ranged from 3.1 to 16.9‰ (Figure 1e). Surface specific removal was 45 g m-2 yr-1, corresponding to a removal efficiency of 27% (Table 1). The PN and NO3-
FIGURE 1. Hydraulic loading rate (a), average daily water temperature (b), NO3- concentration of inlets I and II (c) and the outlet (d), δ15NNO3 of inlet I, inlet II, and the outlet (e), and TN removal of Wetland Boden (f) from January to October 2004.
TABLE 2. Nitrogen Fluxes (mg m-2 d-1) in Wetland Boden (April-October 2004) mg m-2 d-1
NO3- input NO3- output change of the NO3- stock NO3- uptake into PNau plankton duckweed sediment trap a sediment trap, allochthonous denitrification b nitrification b nitrification supplied with NH4+ from mineralized DON c allochthonous NH4+ d benthic NH4+ e
in out dNO3-/dt up PNplankton PNduckweed PNtrap,au PNtrap,al deni nitr nitrDON nitrNH4,al nitrNH4,benthic
Apr 03/22-04/27
May 04/27-05/26
Jun 05/26-06/22
Jul 06/22-07/20
Aug 07/20-08/24
Sep 08/24-09/22
250 163 -31
266 104 -48
374 240 -14
20 27 -8
5 1 -1
22 1 0
3 1 -1
8 0 0 26 129 19
13 15 61 27 222 101
108 54 77 112 119 210
-11 15 116 10 1 120
1 -3 51 9 6 50
7 -75 199 9 5 115
-1 0 56 0 4 56
28 -9 0
77 -19 43
88 -8 130
-13 -44 177
17 -61 94
-6 5 116
-14 -14 84
a Assuming complete settling of PN in the wetland (PN ) PN al al trap,al), PNtrap,au was calculated from PNtrap ) PNtrap,au + PNtrap,al. eqs 1 and 2. c dDON/dt ) in - out - nitrDON. d dNH4/dt ) in - out - nitrNH4,al. e nitr ) nitrDON + nitrNH4,al + nitrNH4,benthic.
load decreased by 42% and 29%, respectively, whereas the NH4+ output balanced the NH4+ input. TN removal varied seasonally and interannually: In 2002, 41% of the TN load, but in 2003 only 16% of the TN load, was removed in the wetland. In 2004, TN removal was -35 mg m-2 d-1 (-4%) in January and February, 164 mg m-2 d-1 (36%) from March to June, and -17 mg m-2 d-1 (-57%) from July to October (Figure 1f). Detailed N fluxes during the isotope investigation from April to October 2004 are shown in Table 2. Nitrogen removal exceeded phosphorus (P) removal 41 fold (18). Wetland Chemistry. According to the temperature profiles (Figure 2a), the wetland’s water column was completely mixed in spring and autumn 2004. During the low water discharge period from June to August, the mixing depth was limited to the upper 1 m. During turnover in April and May, O2 concentration in the water column exceeded 13 mg L-1
b
Oct 09/22-10/18
Calculated from
due to intense photosynthesis; in June, it decreased to values below detection limit right above the sediment (Figure 2d). Nitrate concentration declined from approximately 3.00 mg L-1 in March to 0.10 mg L-1 in June (Figure 2b). Simultaneously, δ15NNO3 decreased from 16.8‰ in March and April to 12.1‰ in May and averaged 3.9‰ after July (Figure 2e). Ammonium concentration was smaller than 0.10 mg L-1 in March and April but was up to 10.0 mg L-1 above the sediment from June to August (Figure 2c). δ15NNH4 ranged from 9.7 to 12.5‰ (Figure 2f). Duckweed covering the water surface from May to August (Figure 1b) reached its maximum density of 280 g dwt m-2 in July. Its N content declined from 50.4 mg g-1 dwt in May to 16.0 mg g-1 dwt in August. Simultaneously, its δ15N decreased from 14.4‰ in June to 8.2‰ in August. Sedimentation rate in the wetland ranged from 4 to 22 g dwt m-2 d-1. VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3315
FIGURE 2. Seasonal variation of the water column chemistry in Wetland Boden in 2004: Water temperature (a), NO3- concentration (b), NH4+ concentration (c), oxygen concentration (d), δ15NNO3 (e), and δ15NNH4 (f) plotted versus depth. The sediment trap material had a minimum N content of 6.4 mg g-1 dwt in March and a maximum N content of 17.1 mg g-1 dwt in August. δ15N of the sediment trap material decreased from 10.2‰ in April to 5.3‰ in December. At the bottom of the main pond, a sediment layer with a mean N content of 2.5 mg g-1 dwt and a δ15N of 5.5‰ ((0.2) accumulated during the 2.5 years. The δ15N of the sediment collected in the sedimentation pond at inlet I (5.1‰ (0.3) reflected the δ15N of allochthonous PN. The N:P ratio of the wetland’s sediment was 2.3. Laboratory Experiments. Denitrification rates followed a Michaelis-Menten equation relative to the NO3- concentration (r2: 0.961). They increased in proportion to the NO3concentration with an initial slope of 0.13 m d-1 and converged on a maximum of 690 mg m-2 d-1 ((70). The half saturation constant was 5.47 mg L-1 (Figure 3).
Discussion Nitrogen Removal. The TN removal of 45 g m-2 yr-1 (27%) observed in Wetland Boden is within the range reported for wetlands (19) and lakes (20) that reduce the N load from nonpoint source pollution by an average of 27% (see Supporting Information). The removed N is either accumulated in the sediments or eliminated by denitrification, whereas P removal is only due to sediment accumulation. Given that 41 times more N than P was removed in Wetland Boden, but only 2.3 times more N than P accumulated in the sediment, only 6% (3 g m-2 yr-1) of the removed N was stored in the sediment, whereas 94% (42 g m-2 yr-1) was denitrified and lost to the atmosphere. TN removal in Wetland Boden was not constant throughout the studied period as shown in Figure 1f. Low water 3316
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 10, 2006
FIGURE 3. Denitrification as determined in laboratory experiments (filled circles) and calculated for Wetland Boden according to the isotope mass balance approach (open triangles) plotted versus NO3concentration. temperature and short water residence time limiting microbial activity and plankton assimilation accounted for the low or even negative TN removal during winter 2004 but could not explain the decline of the absolute and percentage TN removal during summer. Therefore, N transformations were quantified monthly with the isotope mass balance discussed below. Nitrate Isotope Signature To Identify Nitrogen Transformations. Denitrifying bacteria discriminate against the heavier nitrogen isotope, if a sufficient amount of NO3- is available. Thus, during denitrification the microbially produced N2 is depleted of 15N, whereas the remaining NO3-
FIGURE 4. Isotope mass balance model for N removal and recycling in Wetland Boden. N fluxes (arrows), N reservoirs (ovals), and their isotopic signature (‰) as measured or estimated (*) in Wetland Boden from March to June (a) and July to October (b), 2004. (**For water column denitrification a fractionation factor of -24‰ was assumed.) becomes enriched with 15N. Fractionation observed in wetlands (: -2.5‰), however, was far lower than determined in laboratory experiments (: -17 to -29‰) (9). In Wetland Boden, δ15NNO3 remained constant while the NO3- concentration decreased between March and April (Figure 2b, e). In addition, δ15NNO3 decreased until June, and the outlet δ15NNO3 did not exceed those of the inlets (Figure 1e). Two factors summarized in Figure 4 could account for this pattern: (i) As denitrifying enzyme activity requires an O2 concentration below 0.2 mg L-1 (21) and the water column was oxic from March to May, denitrification in spring was limited to the anoxic sediment layer (Figure 4a). Benthic denitrification acts as a perfect sink since nitrate diffusing across the water-sediment interface is completely consumed irrespective of its isotopic composition (22, 23). Hence, denitrification during spring did not produce isotope fractionation. (ii) For the further decrease of δ15NNO3 between May and July, an additional process yielding 15N-depleted NO3- is required. The fixation of atmospheric N2 (δ15N: 0‰) was of minor importance, as an investigation of the wetland’s phytoplankton revealed. Therefore, the light NO3- most likely originated from NH4+ nitrification. As the NH4+ input via the inlets was negligible, additional NH4+ must have been supplied by mineralization of sediment organic matter and DON (Table 2). Because benthic mineralization has been shown to lack fractionation (17, 22), the isotopic composition of NH4+ released from the sediments was similar to the sediment PN (5.5‰). In spring, regardless of these NH4+ supplies, the NH4+ concentration in the water column remained low, indicating complete nitrification. Hence, the isotope signature of NO3- and the parent PN were identical (Figure 4a). In summer, in contrast, the average water column δ15NNH4 of 10‰ and the δ15NNO3 of 3.9‰ indicate that only part of the benthic NH4+ was nitrified, resulting in heavier NH4+ and lighter NO3- (Figure 4b). When interpreting the isotopic signature of NH4+, it must be kept in mind that part of the DON from the drainage inlets may also have contributed to the production of NH4+. In spring DON input exceeded its output by 48%; in summer DON input and output were nearly balanced (Table 2). Considering the inlet DON as an intermediate product of allochthonous PN mineralization to NH4+, δ15NNH4 and δ15NDON must be closely linked to the δ15NPN,al of 5.1‰. With
an enrichment factor of -2‰ determined for PN-derived DON in Lake Michigan and Lake Superior (24), δ15NDON and δ15NNH4 in Wetland Boden would be in the range of, or below, 3.1‰. On the other hand, the δ15N of NH4+ released during decomposition of autochthonous PN may have exceeded the average sediment δ15N of 5.5‰, as the sediment composed of degraded material was significantly lighter than the duckweed (8.2-14.4‰) and the settling material (5.310.2‰) collected in the sediment trap. Part of this discrepancy can be explained by the settling of allochthonous PN (5.1‰), which contributed 55% to the total PN deposition during spring 2004, but only 7% during low loading in summer. The accumulation of light PN in the sediments may as well be due to the preferential release of heavy N under anoxic conditions, which has been described in field and laboratory studies (25-27). These findings, which are in contrast to general expectations, were attributed to the selective decomposition of 15N-enriched PN (e.g., proteins) (25) or bacterial growth on the decomposing material (26, 27). Quantitative Assessment of Nitrogen Transformations. Solving eqs 1 and 2 with the isotope signatures summarized in Figure 4 yields an average denitrification rate of 155 mg m-2 d-1 from March to June 2004 and 4 mg m-2 d-1 from July to October (Table 2). The annual denitrification amounted to 39 g m-2 yr-1, if the spring denitrification rate was applied to all months with high NO3- and oxygen concentrations (November-June). It matched well with the denitrification rate of 42 g m-2 yr-1 determined as the difference between the 2.5-yr N removal and N accumulation in the sediment. Hence, coupled nitrification-denitrification in the sediment, which does not imprint on the δ15NNO3 in the water column and which was not assessed with the isotope approach, seemed to be of minor importance. Nitrification determined from the isotope mass balance was on average 91 mg m-2 d-1 and mostly fueled by NH4+ released from the sediment (Table 2). In spring, DON mineralization may have contributed 49 mg m-2 d-1 to the NH4+ production. Benthic NH4+ release calculated as the difference between nitrification and NH4+ supply from DON ammonification or allochthonous NH4+ ranged from 43 to 177 mg m-2 d-1. The benthic NH4+ release from June to August (131 mg m-2 d-1) agreed remarkably well with the turbulent VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3317
diffusive NH4+ flux (130 mg m-2 d-1) independently calculated from NH4+ concentration gradients in the water column of the wetland (Figure 2c). Denitrification versus Nitrogen Recycling. The average 2.5-yr denitrification rate (42 g m-2 yr-1) was at the upper level of the rates reported for lake, river and estuary sediments (21), but similar to values observed in wetlands (28). Benthic denitrification rates in Wetland Boden increased with increasing NO3- concentration in the water column as the laboratory experiments showed (Figure 3). This indicates that benthic denitrification is controlled by the diffusive NO3supply to the sediment which increases in proportion to the NO3- concentration in the sediment overlying water. In accordance with that, high denitrification rates in spring 2004 were linked to high external NO3- loading and intense nitrification in the water column (Table 2). In contrast, low denitrification rates were observed in summer when NO3concentrations declined because external NO3- load decreased and biomass production outcompeted denitrification. The intense duckweed production provided only shortterm N sink and fed internal N recycling. The N content of the settling matter mainly composed of duckweed (12.2 mg g-1 dwt) was significantly higher than the N content of the sediment (2.5 mg g-1 dwt). This indicates that at least 80% of the settled N was released during benthic mineralization. Due to little supportive tissues and low C:N ratios, floating leaved duckweed and phytoplankton are easily degradable and have been shown to lose up to 75% of their initial biomass within 10 weeks of decomposition (29, 30). Benthic release rates and the N content of the remaining organic sediment in Wetland Boden were similar to those of planktondominated lakes (31-34). In contrast, the typical N content reported for peatland soils (35, 36) was approximately 5-10 times higher than in Wetland Boden, suggesting that rooting macrophytes lead to a better N conservation and long-term storage than duckweed and phytoplankton. As the burial of N in the sediment was of minor importance for long-term N removal, denitrification governed N-removal efficiency in Wetland Boden. Two contrasting situations developed (Figure 4), which are typical for wetlands with a variable water discharge and N-loading regime: TN removal was high when the NH4+ released from the sediment was nitrified in the oxic water and subsequently denitrified in the anoxic sediment. However, N removal was close to zero during stagnant periods, when NH4+ accumulated in the anoxic water layer and photoautotrophic plants intensively assimilated N. To improve the N-removal efficiency of wetlands for treating nonpoint source pollution it is essential to minimize internal N recycling and to maximize denitrification losses. As benthic denitrification depends on high nitrate concentrations (Figure 3), the water layer above the anoxic sediments has to remain oxic to guarantee complete nitrification of NH4+. Keeping the wetland shallow can prevent stratification. Harvesting of floating leaved plants stimulates phytoplankton growth and oxygen production and further adds to the removal of the recyclable biomass N pool. However, complete biomass and sediment removal should be avoided to provide sufficient organic carbon substrate for denitrification. Our study has shown that isotopic analysis of the relevant N pools can help to disentangle complex and variable nitrogen elimination pathways and help to design management strategies for such artificial wetlands.
Acknowledgments We thank Carsten Schubert, Dan McGinnes, and Tom Gonser for proofreading and the anonymous reviewers for their valuable comments on the manuscript. Thanks to Ruth Stierli, Elisa Pina, and Toni Mares for assistance in the laboratory. 3318
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 10, 2006
Supporting Information Available A map of the study site and the correlation between surface specific N input and removal of Wetland Boden and other lakes and wetlands. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Howarth, R. W.; Billen, G.; Swaney, D.; Townsend, A.; Jaworski, N.; Lajtha, K.; Downing, J. A.; Elmgren, R.; Caraco, N.; Jordan, T.; Berendse, F.; Freney, J.; Kudeyarov, V.; Murdoch, P.; Zhu, Z. L. Regional nitrogen budgets and riverine N & P fluxes for the drainages to the North Atlantic Ocean: Natural and human influences. Biogeochemistry 1996, 35, 75-139. (2) Galloway, J. N.; Schlesinger, W. H.; Il, H. L.; Michaels, A.; Schnoor, J. L. Nitrogen fixation: Anthropogenic enhancement-environmental response. Global Biogeochem. Cycles 1995, 9, 235-252. (3) Cloern, J. E. Our evolving conceptual model of the coastal eutrophication problem. Mar. Ecol.: Prog. Ser. 2001, 210, 223253. (4) Paris Commission (PARCOM). Recommendation 88/2 of 17 June 1988 on the Reduction in Inputs of Nutrients to the Paris Convention Area; 1988. (5) Fisher, J.; Acreman, M. C. Wetland nutrient removal: a review of the evidence. Hydrology Earth Syst. Sci. 2004, 8, 673-685. (6) Johnston, C. A. Sediment and nutrient retention by freshwater wetlands: effects on surface-water quality. Crit. Rev. Environ. Control 1991, 21, 491-565. (7) Singleton, M. J.; Woods, K. N.; Conrad, M. E.; Depaolo, D. J.; Dresel, P. E. Tracking sources of unsaturated zone and groundwater nitrate contamination using nitrogen and oxygen stable isotopes at the Hanford Site, Washington. Environ. Sci. Technol. 2005, 39, 539-548. (8) Mengis, M.; Schiff, S. L.; Harris, M.; English, M. C.; Aravena, R.; Elgood, R. J.; MacLean, A. Multiple geochemical and isotopic approaches for assessing groundwater NO3 - elimination in a riparian zone. Ground Water 1999, 37, 448-457. (9) Lund, L. J.; Horne, A. J.; Williams, A. E. Estimating denitrification in a large constructed wetland using stable nitrogen isotope ratios. Ecol. Eng. 2000, 14, 67-76. (10) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry, 2nd ed.; John Wiley & Sons: New Jersey, 2003. (11) DEV German Standard Methods for the Examination of Water, Wastewater and Sludge (in German); Wiley VCH: Weinheim, 2002. (12) Ebina, J.; Tsutsui, T.; Shirai, T. Simultaneous determination of total nitrogen and total phosphorus in water using peroxodisulfate oxidation. Water Res. 1983, 17, 1721-1726. (13) Kendall, C.; Caldwell, E. A. Fundamentals of Isotope Geochemistry. In Isotope Tracers in Catchment Hydrology; Kendall, C., McDonnell, J. J., Eds.; Elsevier: Amsterdam, 1998; pp 51-86. (14) Silva, S. R.; Kendall, C.; Wilkison, D. H.; Ziegler, A. C.; Chang, C. C. Y.; Avanzino, R. J. A new method for collection of nitrate from fresh water and the analysis of nitrogen and oxygen isotope ratios. J. Hydrol. 2000, 228, 22-36. (15) Lehmann, M. F.; Bernasconi, S. M.; McKenzie, J. A. A method for the extraction of ammonium from freshwaters for nitrogen isotope analysis. Anal. Chem. 2001, 73, 4717-4721. (16) Kroeze, C.; Seitzinger, S. P. Nitrogen inputs to rivers, estuaries and continental shelves and related nitrous oxide emissions in 1990 and 2050: a global model. Nutr. Cycling Agroecosyst. 1998, 52, 195-212. (17) Kendall, C.; Doctor, D. H. Stable Isotope Applications in Hydrologic Studies. In Treatise on Geochemistry; Holland, H. D., Turekian, K. K., Eds.; Surface and Groundwater, Weathering, and Soils, Vol. 5; Elsevier: Amsterdam, 2004; pp 319-364. (18) Reinhardt, M.; Ga¨chter, R.; Mu ¨ller, B. Small constructed wetlands - contributing to the restoration of the Central Swiss lakes? (in German). Manuscript submitted. (19) Mitsch, W. J.; Gosselink, J. G. Wetlands, 3rd ed.; John Wiley: New York, 2000. (20) Steingruber, S. Nutrient transformations in a wetland pond. Ph.D. Thesis, Swiss Federal Institute of Technology, Zu¨rich, 2001. (21) Seitzinger, S. P. Denitrification in freshwater and coastal marine ecosystems - ecological and geochemical significance. Limnol. Oceanogr. 1988, 33, 702-724. (22) Brandes, J. A.; Devol, A. H. Isotopic fractionation of oxygen and nitrogen in coastal marine sediments. Geochim. Cosmochim. Acta 1997, 61, 1793-1801.
(23) Sebilo, M.; Billen, G.; Grably, M.; Mariotti, A. Isotopic composition of nitrate-nitrogen as a marker of riparian and benthic denitrification at the scale of the whole Seine River system. Biogeochemistry 2003, 63, 35-51. (24) Feuerstein, T. P.; Ostrom, P. H.; Ostrom, N. E. Isotopic biogeochemistry of dissolved organic nitrogen: A new technique and application. Org. Geochem. 1997, 27, 363-370. (25) Altabet, M. A.; Deuser, W. G.; Honjo, S.; Stienen, C. Seasonal and depth-related changes in the source of sinking particles in the North Atlantic. Nature 1991, 354, 136-139. (26) Libes, S. M.; Deuser, W. G. The isotope geochemistry of particulate nitrogen in the Peru Upwelling Area and the Gulf of Maine. Deep-Sea Res. 1988, 35, 517-533. (27) Lehmann, M. F.; Bernasconi, S. M.; Barbieri, A.; McKenzie, J. A. Preservation of organic matter and alteration of its carbon and nitrogen isotope composition during simulated and in situ early sedimentary diagenesis. Geochim. Cosmochim. Acta 2002, 66, 3573-3584. (28) Xue, Y.; Kovacic, D. A.; David, M. B.; Gentry, L. E.; Mulvaney, R. L.; Lindau, C. W. In situ measurements of denitrification in constructed wetlands. J. Environ. Qual. 1999, 28, 263-269. (29) Hunter, R. D. Changes in carbon and nitrogen content during decomposition of three macrophytes in freshwater and marine environments. Hydrobiologia 1976, 51, 119-128. (30) Ho¨hener, P. The Nitrogen Balance of Lakes - a Case Study of Lake Sempach (in German). Ph.D. Thesis, Swiss Federal Institute of Technology Zurich, 1990. (31) Mu ¨ ller, B.; Ma¨rki, M.; Dinkel, C.; Stierli, R.; Wehrli, B. In Situ Measurements in Lake Sediments Using Ion-Selective Electrodes
(32)
(33)
(34)
(35)
(36)
with a Profiling Lander System. In Environmental Electrochemistry: Analyses of Trace Element Biogeochemistry; Taillefert, M., Rozan, T. F., Eds.; American Chemical Society: Washington, DC, 2002; pp 126-143. Ho¨hener, P.; Ga¨chter, R. Nitrogen cycling across the sedimentwater interface in an eutrophic, artificially oxygenated lake. Aquat. Sci. 1994, 56, 115-132. Friedrich, J.; Dinkel, C.; Grieder, E.; Radan, S.; Secrieru, D.; Steingruber, S.; Wehrli, B. Nutrient uptake and benthic regeneration in Danube Delta Lakes. Biogeochemistry 2003, 64, 373398. Ga¨chter, R.; Wehrli, B. Ten years of artificial mixing and oxygenation: No effect on the internal phosphorus loading of two eutrophic lakes. Environ. Sci. Technol. 1998, 32, 36593665. Bridgham, S. D.; Updegraff, K.; Pastor, J. Carbon, nitrogen, and phosphorus mineralization in northern wetlands. Ecology 1998, 79, 1545-1561. Richardson, C. J.; Tilton, D. L.; Kadlec, J. A.; Chamie, J. P. M.; Wentz, W. A. Nutrient dynamics of northern wetland ecosystems. In Freshwater Wetlands; Good, R. E., Whigham, D. F., Simpson, R. L., Eds.; Academic Press: New York, 1978; pp 217-241.
Received for review November 29, 2005. Revised manuscript received March 10, 2006. Accepted March 13, 2006. ES052393D
VOL. 40, NO. 10, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3319
